Outcome research has shown that social influence-based alcohol prevention programs can be, and generally are, efficacious. However, because the question of whether a program works has been paramount in most prevention studies to date, much less is known empirically about how social influence-based prevention programs work and how they can be made to work better. The main goal of this proposal is to increase the precision of empirical evidence as well as advance theoretical understanding, of social influence-based alcohol abuse prevention interventions. This will be accomplished using a theory-driven evaluation approach, the latest methods for tracing processes that lead to alcohol prevention effects, and by examining under what conditions and for whom prevention programs optimally prevent the onset of adolescent alcohol use, alcohol misuse, and alcohol- related problems. A very unique and rich longitudinal data set containing detailed multitrait-multimethod measures from a large alcohol prevention trial, the Adolescent Alcohol Prevention Trial, will be used to carry out the proposed analyses. Further, an emerging analysis technology, Confirmatory Factor Analysis of Multitrait-Multimethod (MTMM) data, will be employed to understand, and if necessary, control for response or mono- method bias in alcohol-related measures.